mark

2.5K posts

mark banner
mark

mark

@thisisnotmark_

AI/ML Research Lead @ NASA | Founder @ Mosaic Voice AAC. Views my own not my employer’s.

انضم Şubat 2015
1.4K يتبع538 المتابعون
تغريدة مثبتة
mark
mark@thisisnotmark_·
will be presenting my paper “Life, Machine Learning, and the Search for Habitability: Predicting Biosignature Fluxes for the Habitable Worlds Observatory” tomorrow at #AAAI. if anyone at AAAI wants to meet for coffee let me know! arxiv.org/abs/2601.12557
English
0
0
1
361
mark أُعيد تغريده
Anish Moonka
Anish Moonka@anishmoonka·
NASA has 32 cameras on the Artemis II spacecraft. The top science priority during the Moon flyby was the four astronauts looking out the window and talking about what they saw. NASA's lunar science lead confirmed it. What the crew says out loud about the Moon's surface matters more to the science team than anything the cameras capture. NASA trained this crew in Iceland's volcanic highlands and at an impact crater in Labrador, Canada, teaching them to read rock textures and spot geological details at 25,000 mph. There's a reason NASA trusts human eyes over cameras. In 1972, Apollo 17 astronaut Harrison Schmitt was walking near a small crater called Shorty when he scuffed the dirt with his boot. The soil underneath was orange. Schmitt was the only trained geologist to ever walk on the Moon, and he got so excited he blurred most of his own photos. That orange soil turned out to be tiny glass beads from a volcanic eruption 3.64 billion years ago, one of the biggest finds of the entire Apollo program. A boot and a pair of trained eyes caught what no camera did. For this flyby, NASA sent the crew a final list of 30 surface targets. They killed all the cabin lights to cut window reflections. They worked in pairs, rotating every 55 to 85 minutes, calling out craters and lava flows while scientists at Johnson Space Center analyzed everything in real time. Pilot Victor Glover reported that the Moon's south pole, where NASA wants to land astronauts by 2028, looked "more jagged" than the north with much steeper terrain. One observation from a human eye at 4,070 miles could shape where the next crew touches down. At 6:44 PM Eastern, Orion slipped behind the far side and went radio silent for 40 minutes. Four people, completely cut off from every other human alive, the Moon blocking every signal back to Earth. The last time humans experienced that was December 1972. They broke the all-time distance record on the way. Apollo 13 held it for 56 years at 248,655 miles from Earth. Artemis II passed that mark and kept going to 252,760. Jim Lovell, who commanded Apollo 13 and held that record his whole life, died last August at 97, eight months before these four beat it. Before he died, Lovell recorded a message for the crew. "Welcome to my old neighborhood," he told them. "Don't forget to enjoy the view." The crew named two craters during the flyby. One for their spacecraft, Integrity. The other, Carroll, for Commander Reid Wiseman's late wife, a nurse who cared for newborns and died of cancer in 2020 at 46. Wiseman has raised their two daughters alone since. When Jeremy Hansen read the name to Mission Control, his voice broke. The crew hugged. Wiseman and Koch wiped tears. Then they got back to work, because they still had hours of Moon left to map with their eyes.
NASA@NASA

LIVE: Watch with us as the Artemis II astronauts make their closest approach to the Moon, traveling farther from Earth than ever before. twitter.com/i/broadcasts/1…

English
23
715
5.6K
577.5K
mark أُعيد تغريده
Jeremy
Jeremy@ManaByte·
Thank you to everyone at @NASA for showing the world what a space program can still do when it chooses to inspire. The four heroes of Artemis II have flown around the moon and are on their way home and none of this happens without the thousands of people who give everything to make missions like this real. You’ve lit a spark for a new generation of explorers who are seeing what’s possible again. And thank you to @NASAAdmin for guiding this new era of human spaceflight beyond Earth with steady leadership and a clear vision. This is history and you helped bring it back to life.
English
2
78
607
7.4K
mark أُعيد تغريده
NASA Administrator Jared Isaacman
Artemis II has reached its maximum distance from Earth. On the far side of the Moon, 252,756 miles away, Reid, Victor, Christina, and Jeremy have now traveled farther from Earth than any humans in history and now begin their journey home. Before they left, they said they hoped this mission would be forgotten, but it will be remembered as the moment people started to believe that America can once again do the near-impossible and change the world. Congratulations to this incredible crew and the entire NASA team, our international and commercial partners, but this mission isn’t over until they’re under safe parachutes, splashing down into the Pacific.
English
1.3K
7.1K
53.5K
6.7M
mark أُعيد تغريده
NASA Artemis
NASA Artemis@NASAArtemis·
"We will always choose each other." Mission control has reacquired signal with the Artemis II crew after the mission’s planned loss of signal. Our astronauts are once again using the Deep Space Network to keep conversation and science data flowing between space and Earth.
English
192
4K
19.8K
445.7K
mark
mark@thisisnotmark_·
my heart is so happy. sitting here watching this live stream all day is so soothing. I feel like the world is coming together on this and is undivided in a way it hasn’t felt like in many years. today, we are all human. today, the borders are blurred and we all cheer for mankind together 🚀🚀🚀
English
0
0
0
30
mark أُعيد تغريده
Luc Riesbeck ♻️🛰
Luc Riesbeck ♻️🛰@LucRiesbeck·
This is one of those days where I am deeply, deeply proud to work at NASA. 🚀🌕
English
6
6
233
3.3K
mark أُعيد تغريده
Thomas Ricouard
Thomas Ricouard@Dimillian·
🚀 ❤️ 🌕
mark@thisisnotmark_

@Dimillian sitting here in my office at NASA using codex right now. will try my hardest not to ship slop to space 🚀

ART
1
1
17
3.4K
mark
mark@thisisnotmark_·
@Dimillian sitting here in my office at NASA using codex right now. will try my hardest not to ship slop to space 🚀
English
3
1
19
3.8K
Thomas Ricouard
Thomas Ricouard@Dimillian·
We’re going to the moon and you’re shipping slop, use /review
English
13
1
86
6K
mark
mark@thisisnotmark_·
oops accidentally sent humans the farthest they’ve ever been to space @NASA
English
0
0
0
59
anirudh bv
anirudh bv@anirudhbv_ce·
@thisisnotmark_ Please try it out and let me know how it goes, Mark! Always on the lookout for feedback and suggestions :)
English
1
0
2
147
mark
mark@thisisnotmark_·
big if true. excited to try this out! great job
anirudh bv@anirudhbv_ce

pip install turboquant-gpu 5.02x KV cache compression for ANY GPU (RTX, H100, A100, B200) - works over @huggingface transformers - dead-simple API: compress + generate in 3 lines - 3-bit Lloyd-Max fused KV compression (0.98 cosine similarity) - outperforms MXFP4 (3.76x) and NVFP4 (3.56x) on compression Ran Mistral-7B: 1,408 KB → 275 KB KV cache (5.02x) Quickstart: github.com/DevTechJr/turb… Written in cuTile (CUDA 12, 13) with PyTorch fallbacks

English
1
0
3
269
mark
mark@thisisnotmark_·
@NASA explain yourself
English
0
0
0
22
mark
mark@thisisnotmark_·
> 4 people > 3 guys 1 girl > venturing further in space than anyone ever before > main guy's name is Reid (pronounced "Reed") 🧐
mark tweet mediamark tweet media
English
1
0
0
60
mark أُعيد تغريده
Ellie Sleightholm
Ellie Sleightholm@elsleightholm·
Today is the day! Where will you be watching from? Credit: @NASA
Ellie Sleightholm tweet media
English
48
36
639
16K
mark
mark@thisisnotmark_·
those who are building agents for open endedness in scientific research, I'm curious about how you're building things out. I have settled on a pretty decent pipeline for myself, but would be happy to talk and share notes! hope the algo reaches you
English
0
0
0
17
mark
mark@thisisnotmark_·
I'm building my own platform for this right now. my work is at the intersection of building AI/ML in the space and earth sciences, so naturally I thought why not make the science part agentic as well. hardest part right now is getting compute cost for frontier models. OSS models only go so far and are generally slower with my hardware, but make open endedness feasible. other companies are forming around AI co-scientists on a more robust scale. @SakanaAILabs in particular is one I am very excited about
English
1
0
1
47
albina
albina@enjojoyy·
I want to see the success of coding agents replicated in research and science
English
19
6
60
2.4K
mark
mark@thisisnotmark_·
@BaffledCentrist @sircalebhammer I love his interviews. they're so calming I feel like my brain can process the actual information from the interview more. can think deeply about what I resonate with also your handle says you're a centrist. doesn't that mean you're also naturally inclined to see both sides lol
English
0
0
5
230